Frontiers in Business, Economics and Management ISSN: 2766-824X | Vol. 5, No. 3, 2022 301 Design of Quantitative Investment Strategy Based on Domestic Chip Industry Lingtao Qian, Tianyu Chen, Xiuchun Wang Anhui University of Finance and Economics, Bengbu, Anhui, 233030, China Abstract: Select 59 concept stocks in the chip industry for quantitative analysis, use principal component analysis method and k-means clustering to process our financial indicators; use historical simulation method, variance-covariance method and Monte Carlo simulation method to calculate the VaR value of each stock for effective risk management. Based on Markowitz theory, the effective frontier of five stocks is constructed, combined with the utility of investors in different risk aversion levels, the best portfolio is selected to obtain the best portfolio of concept stocks in the chip field, using the Sharp ratio and CV ratio for performance evaluation, and the corresponding investment suggestions are given. Keywords: Chip industry, Quantitative investment, Principal component analysis, Portfolio, Performance evaluation. 1. Foreword A new round of global scientific and technological innovation in the new era has been started. Semiconductor chips, as the foundation of "5G + AI", are the core cornerstone of high-end manufacturing industry. At the same time, due to the development of the high-tech market, the demand for chips is increasing. From the micro perspective, the investment analysis of chip industry concept stocks, on the one hand, helps investors to optimize capital distribution and build investment portfolio, enhance industry competitiveness; on the other hand, it helps to help enterprise investment decision makers to make a correct and reasonable judgment, and obtain higher investment returns. From a macro point of view, the quantitative investment analysis of the chip industry concept stocks is conducive to finding the deficiencies of China's existing asset evaluation system and promoting the improvement and improvement of the relevant asset evaluation system. In general, the investment value analysis of the chip industry not only meets the needs of investment decision-makers, but also meets the needs of shareholders to obtain income. 2. Quantitative Stock Selection This study adopts quantitative stock selection, selecting a few highly growth and highly profitable stocks from the stocks of all chip semiconductor sectors. Through the analysis of the fundamentals, growth, and profitability of individual stocks, select the best growth stocks, and build the stock portfolio. By using the quantitative investment strategy, the selected stocks are more accurate and systematic. 2.1. Alternative Stock Pool In order to make the screening accuracy higher, 19 stocks in all chip semiconductor sectors were preliminarily screened, and 19 stocks with return on equity (ROE) greater than 8% in 2021 were selected as the alternative stock pool of stock selection for this investment. Table 1. 19 Stocks with return on equity greater than 8%:% Unigroup Micro Zhuo Sheng Zhongying Electronics, Zhongke Shuguang, the Great Wall of China ROE 9.69 29.19 19.59 13.73 13.44 Zhaoyi innovation, Tai Chi industry, Shengong shares, Shengbang shares Nasida ROE 11.61 9.11 21.33 15.78 12.9 Leon Micro Lexin Technology, Montage Technology, Kangtuo Infrared, Jingsheng Electromechanical ROE 8.47 9.85 12.73 9.94 14.01 Jiangfeng Electronics, Huiding technology and Taihang Jin Technology ROE 9.4 35.99 14.55 11.43 2.2. Select Financial Indicators When selecting the growth stocks, the financial indicators are mainly selected from the three aspects of growth ability, profitability and solvency to analyze the listed companies. 2.3. Principal Component Analysis There is a certain correlation between the various financial indicators of the stocks, so the information reflected through the index data overlaps to a certain extent. In order to analyze the problem comprehensively and accurately, the study chose 302 the principal component analysis method to reduce the index dimension, analyze the correlation between the indexes, and select the stocks with high growth and high profitability combined with k-means clustering. Table 2. Results of the principal component analysis Total variance interpretation ingredient Initial eigenvalue Extract the sum of load squares amount to variance percentage accumulate% amount to variance percentage accumulate% 1 5.323 59.144 59.144 5.323 59.144 59.144 2 1.700 18.889 78.032 1.700 18.889 78.032 3 1.039 11.549 89.581 1.039 11.549 89.581 4 0.526 5.844 95.425 5 0.249 2.768 98.193 6 0.084 0.932 99.125 7 0.054 0.600 99.725 8 0.024 0.261 99.987 9 0.001 0.013 100.000 Judging from the principal component feature root and contribution rate, the first three feature roots are 5.323,1.700 and 1.039, respectively, and the cumulative variance contribution rate of the first three principal components reached 89.581%, covering most of the information. It shows that the first three principal components can represent the first 9 indicators to analyze whether the stocks have high profitability and high growth, and calculates the scores and comprehensive scores of the three principal components of the 19 stocks. The calculation results are as follows: Table 3. Main component scores and composite scores Securities referred to as Growth ability profitability debt paying ability Comprehensive score ranking Zhuo Sheng micro 1.97335 0.45925 1.97638 1.40440 1 Huiding technology 3.02858 -0.01196 0.08053 1.23893 2 Montage science and technology -0.82612 2.36953 -0.34355 0.48181 3 Lexin Technology -1.04977 1.48790 1.36817 0.43458 4 ShenGong shares 0.68389 1.27197 -2.22871 0.26712 5 ZhaoYi innovation -0.44230 0.77614 0.43667 0.20863 6 Shengbang shares 0.24175 0.34861 -0.14711 0.19677 7 Zhongying electronics 0.39784 0.42127 -0.95706 0.10986 8 Kang Tuo infrared -0.74489 -0.22882 2.19598 0.09347 9 Jingsheng mechanical and electrical -0.09560 -0.05424 -0.53028 -0.17526 10 And tai -0.09065 -0.59561 0.06014 -0.24759 11 Unigroup national micro -0.61188 -0.06499 0.03320 -0.26483 12 Hangjin technology -0.50149 -0.13004 -0.66001 -0.39647 13 the Great Wall of China -0.12739 -0.86385 -0.30298 -0.44291 14 Zhongke dawn 0.01254 -1.17998 -0.13294 -0.46796 15 Jiang Feng electronics -0.61852 -0.66563 0.05138 -0.48950 16 Leon micro -0.63276 -0.56341 -0.51599 -0.58109 17 Tai chi industry -0.51463 -1.02936 -0.13606 -0.62535 18 Nathda -0.08196 -1.74678 -0.24776 -0.74460 19 It can be seen from the comprehensive ranking of stocks, investors with limited funds can choose the top-ranked Zhuoshengwei and Huiding Technology for invest, while investors with sufficient funds can choose to make portfolio investment. However, PCA can only derive rankings between stocks and cannot classify stocks, so a combined clustering analysis can be used to further classify stocks. 2.4. K-means Clustering Investment objects were clustered by k-means clustering to identify appropriate classification criteria. Investors can focus on a particular category in order to further study a particular category of stocks to build an optimal portfolio. 303 Table 4. K-means clustering results Cluster 1 Unigroup Guowei, Zhongying Electronics, Zhongke Shuguang, Great Wall of China, Taiji Industry, Shengbang Shares, Nasida, Leone Wei, Jingsheng Mechanical and Electrical, Jiangfeng Electronics, Heertai, Hangjin Technology Cluster 2 Zhuoshengwei, Kangtuo Infrared, Huiding Technology Cluster 3 Zhaoyi Innovation, Shengong Shares, Lexin Technology, Montage Technology Table 5. K-means cluster analysis Cluster 1 Cluster 2 Cluster 3 Growth ability 0.21856 1.41901 0.40858 Profitability 0.51033 0.07282 1.47639 Debt 0.29046 1.41763 0.19186 According to k average clustering, 19 stocks are divided into three categories, as can be seen from the table the second class of stocks for blue chips, comprehensive performance is better than the other two categories, suitable for long-term investment: the third class stock short-term profitability, but poor growth, suitable for short-term investment: the first class of stocks on each three sides, are poor, for poor performance stocks, not suitable for investment. 2.5. Preliminary Construction of The Stock Portfolio Combine the principal composition analysis and k-mean clustering methods to construct the stock portfolio. The types of stocks selected are: Zhuosheng Micro, Huiding Technology, Montage Technology, Lexin Technology and Kangtuo Infrared to make long-term investment and further study the risks of stocks. 2.6. Risk Management Parameter method, historical simulation method and Monte Carlo simulation method are used to calculate the actual risk value of the selected stocks for effective risk management. According to the obtained stock price, the yield rate of each stock is calculated to obtain the in-risk value of the portfolio under the three methods. Suppose that the value of the portfolio is RMB 100 million respectively. The calculation results are as follows: Table 6. Values of the five stocks Zhuo Sheng micro Huiding technology Montage science and technology Lexin Technology Kang Tuo infrared covariance 0.00001195 0.00000471 0.00000304 0.00001395 0.00000764 variance 0.00015555 0.00015050 0.00013839 0.00015555 0.00012815 β 0.07681404 0.03126237 0.02193265 0.07682504 0.05963223 Expected yield 0.01404410 0.01440986 0.01457958 0.01404510 0.01389261 By the high yield, high risk theory: the higher the expected yield, the higher the underlying risk value faced. ZhuoSheng micro and le xin technology in the selected stock is the largest price volatility, the expected yield is low, to eliminate one of the stock portfolio, again in the risk value calculation, after eliminating stock portfolio in risk decline, may not be very suitable for investment, after the investment construction need to focus on the influence of the two stocks. 3. Portfolio Construction 3.1. The Construction of The Stock Effective Frontier To seek the optimal ratio between assets, this study, based on Markowitz theory, builds an effective frontier of five stocks (generating 50 portfolios with different weight ratios) with MATLAB. The results are as follows: Figure 1. Effective frontier chart of the five stock data Select data from the effective frontier for the following table: 304 Table 7. Effective cutting-edge data for the 5 randomly selected stocks Investment in Zhuosheng Wei Huiding Technology Lanqi technology Lexin technology Contour infrared VaR(α=0.99) VaR(α=0.95) Port Return Port Risk 1 0.000 0.027 0.450 0.000 0.523 18.408 18.408 0.147 0.004 2 0.000 0.078 0.547 0.000 0.375 19.437 19.438 0.169 0.005 3 0.001 0.122 0.648 0.000 0.229 20.403 20.403 0.192 0.006 4 0.002 0.166 0.748 0.000 0.084 21.368 21.368 0.214 0.006 5 0.005 0.268 0.728 0.000 0.000 22.660 22.660 0.236 0.007 6 0.011 0.631 0.358 0.000 0.000 26.134 26.135 0.281 0.010 7 0.013 0.790 0.141 0.056 0.000 26.509 27.787 0.304 0.012 8 0.021 0.947 0.000 0.032 0.000 28.579 29.307 0.326 0.014 9 0.824 0.176 0.000 0.000 0.000 35.840 35.842 1.045 0.168 10 0.899 0.101 0.000 0.000 0.000 36.434 36.436 1.112 0.184 Each allocation scheme is obtained from the MATLAB calculation results as shown in the table above. The calculation results are the same as the analysis results in the third part, and Lexin Technology occupies a relatively small proportion, almost approaching 0. As can be seen from the data in the table, the higher the combination dispersion degree, the lower the risk it bears. 3.2. Determination of the Optimal Portfolio Weight The utility function is used to calculate the utility of 10 different investment schemes for investors, and the optimal portfolio weight is obtained according to the utility maximization: 𝑈 𝐸 𝑟 1 2 𝐴𝜎 E (r) is the expected return rate of the portfolio, the standard deviation of the portfolio, and A is the risk aversion of investors. In this paper, we believe that all investors are risk aversion (A> 0), and the choice of A is 1,1.5,2 and 2.5, respectively, to calculate the investor utility under different A values. The results are as follows: Table 8. Investor Utility at different A values investment program U(A =1) U(A =1.5) U （A =2) U(A =2.5) 1 0.14448135 0.14233278 0.14233274 0.14125837 2 0.16660912 0.16414825 0.16414825 0.16291775 3 0.18870915 0.18589836 0.18589838 0.18449287 4 0.35078625 0.34378625 0.34378623 0.33568523 5 0.23280845 0.22920691 0.22920692 0.22740612 6 0.27619953 0.27109945 0.27109931 0.26854875 7 0.29770453 0.29165912 0.29165968 0.28863625 8 0.31913752 0.31207521 0.31207525 0.30854375 9 0.32026351 0.31713483 0.31703272 0.30654338 10 0.22199946 0.21189535 0.21189464 0.20274653 Through calculation, it can be seen that under the condition of different degrees of risk aversion, the utility value of investment plan 4 is higher than other schemes, so the portfolio weight of investment plan 4 is the optimal portfolio weight, namely 0.2%, top technology 16.6%, 74.8%, technology 0%, kang infrared 8.4% is the optimal portfolio weight. 3.3. Performance Evaluation The performance evaluation results are obtained by collecting the daily closing price of 19 stocks from October 8 to 31, December 31,2021, and using the portfolio obtained with each stock weight of 1 / 19 as the benchmark for performance evaluation; the Sharp and CV ratios are calculated by the best portfolio scheme and compared with the relevant data of the benchmark. Table Sharpe ratios and CV ratios for the 9 5 stocks Combined average yield variance standard deviation Sharpe ratio CV ratio 0.0035 0.0000001 0.0003 3.2360 0.0896 After calculation, it was found that the Sharp ratio of the five stocks was about 3.2360, and the value was higher than the Sharp ratio of 1.597 for the 19 stocks, indicating that the excess return of the portfolio was significantly higher than the benchmark. So the portfolio we chose is good for the investment. Through the CV ratio, the CV ratio of the portfolio in different samples with the same coefficient is 0.0896. The CV ratio of less than the benchmark is 0.1630, indicating that the portfolio takes less risk for each unit of return, so the selected portfolio has relatively good performance. 305 4. Summary The portfolio strategy is mainly based on risk control. Due to the high volatility of concept stocks, it plays a crucial role in stock selection based on financial indicators. Meanwhile, the conclusion shows that the portfolio construction is relatively successful and is useful for practical operations in different industries. Acknowledgment Foundation Item: National Innovation and Entrepreneurship Training Program for College Students (202110378052). References [1] Forest starts a prairie fire. Application of Economic Profit in Investment Decision and Performance Evaluation in Real Estate Industry [J]. China's collective Economy, 2022 (21): 61- 63. [2] He Yuhan. Research on Quantitative Investment Strategies based on Matlab [J]. Business, 2022 (07): 143-146. [3] Wang Gangzhen, Li Wenbo, Zhu Jiaming. A Quantitative Portfolio Strategy based on the AI Industry [J]. Journal of Shenyang University (Social Sciences edition), 2022,24 (01): 44-51. [4] Zou Luxi. Quantitative investment research based on a multifactor model [J]. China Business Theory, 2021 (24): 100- 103. [5] Zhou Hang. Risk assessment and prevention and control strategies of quantitative investment in financial derivatives [J]. Bohai Rim Economic Outlook, 2021 (05): 29-31. [6] Ding Qi. Research on Stock Multifactor Quantitative Investment Strategies Based on Principal Component Analysis [J]. Time Finance, 2020 (17): 74-76.